Table 2. Statistical comparison of model performance.

Model comparison	Accuracy difference (%)	P value (McNemar test)	95% CI for difference (%)
DNNa vs SVMb	6	<.001	3.8‐8.2
DNN vs RFc	4	.003	1.4‐6.6
DNN vs GBMd	2	.04	0.1‐3.9
GBM vs RF	2	.048	0.02‐4
GBM vs SVM	4	.002	1.5‐6.5
RF vs SVM	2	.04	0.1‐3.9

^aDNN: deep neural network.

^bSVM: support vector machine.

^cRF: random forest.

^dGBM: gradient boosting machine.